Structure and Mechanism of Formation of an Important Ion in Doping Control Chad R

International Journal of Mass Spectrometry 247 (2005) 48–54

Structure and mechanism of formation of an important ion in doping control Chad R. Borges a,∗, James Taccogno a, Dennis J. Crouch a, Ly Le b, Thanh N. Truong b a Sports Medicine Research and Testing Laboratory, Department of Pharmacology and Toxicology, University of Utah, 417 Wakara Way, Suite 2111, Salt Lake City, UT 84108, USA b Department of Chemistry, University of Utah, Salt Lake City, UT 84112, USA Received 14 June 2005; received in revised form 15 August 2005; accepted 15 August 2005 Available online 7 October 2005

Abstract An ion with m/z 143 serves as a biomarker that is often continuously monitored in urine samples undergoing screening by electron ionization gas chromatography/mass spectrometry (EI GC/MS) for banned anabolic agents. The ion is known to arise from trimethylsilyl (TMS)-derivatized synthetic 17-hydroxy, 17-methyl steroids. The purpose of this work was to characterize, in detail, the origin(s), structure(s), and mechanism(s) of formation of such ions with m/z 143. High resolution mass spectrometry (HRMS) data revealed the elemental composition of the D-ring derived m/z 143 ion to be C7H15OSi. Analysis of dihydrotestosterone (DHT) and its 2-methyl substituted analog dromostanolone by HRMS revealed that an elementally equivalent ion of m/z 143 could be derived from the A-ring of TMS-derivatized 3-keto-enol steroids demonstrating that an abnormally intense peak in the m/z 143 extracted ion chromatogram of urine samples undergoing screening for banned anabolic agents does not necessarily indicate the presence of a 17-hydroxy, 17-methyl steroid. To gain information on ion structure, breakdown curves for the most abundant product ions of the m/z 143 ion were generated using both native and perdeutero-TMS derivatives, providing structures for second, third, and fourth 2 generation product ions. An EI-mass spectrum of [16,16,17- H3]-DHT (DHT-d3) demonstrated that one of the C-16 hydrogen atoms is removed prior to the formation of an ion that is highly analogous to the ion with m/z 143 strongly suggesting, in accord with all other evidence, one particular fragmentation pathway and resulting product: a resonance stabilized 3-(O-trimethylsilyl)but-1-ene ion. © 2005 Elsevier B.V. All rights reserved.

Keywords: m/z 143; Steroids; Anti-doping; Ion structure

1. Introduction anabolic doping offense. Details of the offense could be sorted out after an initial positive report, but the facile and rapid acqui- One of the primary goals of anti-doping research is to define sition of a positive result based on a general biomarker would biomarkers of doping, i.e., chemical entities that, when detected greatly simplify the analytical process. This idyllic scenario may at or above certain concentrations, are indicative of a doping never become possible, but the dramatic simplification of sample offense by the provider of the urine or blood sample under screening procedures appears to be within the realm of modern analysis. Although these markers are generally the banned sub- technology. stances or their metabolites, classification as such is not a req- For over a decade, the technique of choice for screening uisite condition for their establishment as a doping biomarker. urine samples for banned anabolic agents has been gas chro- In fact, due to the manner in which sporting federation rules matography coupled to electron ionization mass spectrometry are written, ideal biomarkers would be general in nature such (EI GC/MS) [1–3]. Analyses of urine extract samples by GC/MS that detection of a single marker substance would indicate an are typically run in selected ion monitoring (SIM) mode with multiple ion windows to obtain required sensitivity limits [3], but ions produced with a m/z of 143 are continually monitored in ∗ Corresponding author. Tel.: +1 801 587 8092; fax: +1 801 581 5034. every sample by many world anti-doping agency-accredited lab- E-mail address: [email protected] (C.R. Borges). oratories during the course of screening trimethylsilyl (TMS)-

Fig. 1. EI-mass spectrum of TMS-derivatized epimetendiol (7)—a representa- tive 17-hydroxy, 17-methyl synthetic steroid. The peak at m/z 448 represents the molecular ion (M+). The origin, structure, and mechanism of formation of the ion represented by the base peak at m/z 143 are the focus of the work presented here. derivatized urine sample extracts for banned anabolic substances (personal communication with Larry Bowers, United States Anti-Doping Agency Senior Managing Director of Technical and Information Resources).1 A thorough investigation of the structure, molecular origin(s), and mechanism of formation of this ion, however, has not been reported even though work closely related to some aspects of that described here has been Fig. 2. Extracted ion chromatograms of m/z 143 from urine samples ana- carried out [4,5]. It is understood that the ion consistently arises, lyzed by GC/MS. Panel 1 shows a chromatogram from a blank urine sample. Panel 2 shows a chromatogram from a urine sample spiked with 20 ng/ml of frequently as the base mass spectral peak (Fig. 1), from the D-ring of 17-hydroxy, 17-methyl steroids (1) [5,6] a group epimetendiol (7) and 3 -hydroxystanozolol (3). Samples were prepared and extracted as described elsewhere [21], but derivatized with MSTFA/ammonium of steroids that are predominately synthetic in origin. All the iodide/ethanethiol (1000:2:4, v/w/v), and analyzed by GC/MS as described in steroidal sources of this unique ion, however, have yet to be Section 2. established. Thus, the ion at m/z 143 has served as a (imperfect) biomarker for the presence of synthetic steroids: abnormal and intense chromatographic peak(s) in the m/z 143 extracted ion androstan-17␣-methyl-17␤-ol-3-one), and dihydrotestosterone chromatogram of a urine sample are typically subjected to fur- (DHT) (5␣-androstan-17␤-ol-3-one) were obtained from ther investigation. An example of the utility of this practice is Steraloids (Newport, RI, USA). Epimetendiol (17␤- illustrated in Fig. 2. A better understanding of the origin, struc- methyl-5␤-androst-1-ene-3␣,17␣-diol), 3-hydroxystanozolol ture, and mechanism of formation of this ion may provide a (3,17␤-dihydroxy-17␣-methyl-5␣-androstan [3,2-c] pyra- greater insight into its utility as a urinary biomarker of synthetic zole), 13␤,17␣-diethyl-3␣-17␤-dihydroxy-5␣-gonane, and 2 steroid use. [16,16,17- H3]-DHT (DHT-d3) were obtained from the National Analytical Reference Laboratory (New South Wales, Australia). N-Methyl-N-trimethylsilyltriﬂuoroacetamide (MSTFA) and bis(trimethylsilyl)acetamide (BSA) were pur- chased from Regis Technologies, Inc. (Morton Grove, IL, USA). Ammonium iodide and ethanethiol were acquired from Sigma (St. Louis, MO, USA). Bis(trimethylsilyl)acetamide-d18 (BSA-d18) was obtained from Cambridge Isotope Laborato- ries, Inc. (Andover, MA, USA). Perﬂuorokerosene-H (PFK) was obtained from Lancaster Synthesis Inc. (Pelham, NH, USA).

2. Materials and methods 2.2. Sample preparation 2.1. Materials All steroid samples derivatized with MSTFA were prepared at a ﬁnal concentration of 300 ng/␮l in MSTFA/NH I/ethanethiol Ethylestrenol (4-estren-17␣-ethyl-17␤-ol), dromostanolone 4 (1000:2:4, v/w/v) by heating at 75 ◦C for 30 min. The NH I (5␣-androstan-2␣-methyl-17␤-ol-3-one), mestanolone (5␣- 4 and ethanethiol ensure near-complete keto-enol trimethylsilylation (at the low risk of ethyl thio-incorporation into 1 There are no published reports of this practice, however, it is a well known, silylated steroid structures) [7,8], and therefore eliminates well established practice within the anti-doping community. the need for methyl oxime derivatives. Following derivatiza- 50 C.R. Borges et al. / International Journal of Mass Spectrometry 247 (2005) 48–54 tion samples were injected directly into the GC–MS. Sam- reported. PFK was used as an internal calibrant gas. Follow- ples derivatized with BSA and BSA-d18 were brought up ing calibration, analytical values for PFK ions were always + to the same ﬁnal concentration, but heated overnight at found to be within 0.2 ppm mass accuracy including a C4F5 60 ◦C. Keto-enol trimethylsilylation was generally not complete ion with m/z 142.99201, which was baseline resolved from with this reagent and procedure, but this matter is inconse- peaks of nominal mass 143 derived from TMS-derivatized quential as differentially derivatized molecules are separated steroids. by GC. 2.4. Quantum mechanical (density functional theory) 2.3. Gas chromatography/mass spectrometry calculations

Analysis of extracted urine samples by GC/MS was done Theoretical investigations using density functional theory on an Agilent GC-MSD instrument consisting of an Agilent (DFT) at the B3LYP/6-31G(d,p) level were carried out to pro- 6890 GC coupled to an Agilent 5973 Inert MSD (G2579A vide first-principles information on the relative stability of any performance turbo EI MSD model). One-microliter samples ion structure candidates. The B3LYP hybrid DFT method has were injected in split mode (1:20) onto an injector kept at been known to provide rather accurate energetic properties and 280 ◦C. The GC was operated in constant flow mode with vibrational frequencies of molecular species.2 Candidate struc- a starting carrier gas (He) linear velocity of 35 cm/s through tures for the ion with m/z 143 were fully optimized and all a DB-1MS, 30 m × 0.25 mm I.D., 0.1 ␮m film capillary col- possible resonance structures were taken into account. Normal umn. The initial oven temperature was set to 180 ◦C and mode analyses of these species were also performed to confirm without any initial hold time was ramped at 3.3 ◦C/min to their stability and to predict their infrared (IR) spectra for fur- 231 ◦C, followed by an immediate ramp to 310 ◦C and hold for ther possible identification. All calculations were done using the 2 min, during which the oven was again heated to 325 ◦C and Gaussian03 program [9]. held for 1 min. The GC–MS transfer line was kept at 322 ◦C. The ion source was operated in EI mode at 230 ◦C, with a 3. Results and discussion 70 eV filament. The quadrupole was operated at 150 ◦C and scanned from m/z 50–700 at a rate of 2.29 scans/s. The spec- 3.1. HRMS tra from each scan were recorded individually, i.e., without averaging. The TMS derivatives of two 17-hydroxy, 17-methyl steroid Analysis of samples by GC/MS/MS was carried out on a standards–mestanolone (2) and 3-hydroxystanozolol (3) which Varian 3400F GC (equipped with a Finnigan MAT A200S produce m/z 143 ions when subjected to EI conditions were autosampler and a Varian 1077 split/splitless capillary injector) analyzed by direct insertion probe-EI-HRMS in order to estab- connected to a Finnigan MAT TSQ7000 triple quadrupole mass lish the elemental composition of these ions. A related 17- spectrometer. One-microliter injections of samples were made hydroxy, 17-ethyl synthetic steroid, ethylestrenol (4), which in splitless mode into an injector set at 270 ◦C and a pressure of produces an m/z 157 ion under EI conditions, was also analyzed 8 psi (∼35 cm/s linear velocity) of helium. The GC oven, con- by EI-HRMS. Given the known steroid elemental composi- taining a DB-1MS, 30 m × 0.25 mm I.D., 0.25 ␮m film column, tions, established valency rules, and a mass accuracy within was ramped (after an initial 1 min hold) from 100 to 310 ◦Cat 15 ppm (well within the range of the evaluated instrument 20 ◦C/min and held for 2 min. The ion source was operated in EI accuracy), results indicated that the only possible elemental ◦ mode at 230 C, with a 70 eV filament. Argon was used as the composition assignments are C7H15OSi for the ion with m/z collision gas and was set at a pressure of 3.0 mTorr. Collision 143 and C8H17OSi for the ion with m/z 157. Besides pro- energies were varied as described. In MS/MS mode, the instru- viding elemental composition, these assignments provide evi- ment was operated in product ion scan mode, scanning a mass dence for the suspected D-ring origin of the fragment ions range from 12 to 10 m/z units above the selected precursor ion at in question, which is based on the predicted and observed a rate of 2 scans/s. One precursor ion per run was isolated indi- 14 m/z unit difference between the 17-methyl and 17-ethyl vidually for collision-induced dissociation (CID) with a 1 m/z steroids. unit selection window. A calculation of the number of rings and double bonds in Analyses of samples by direct insertion probe, high resolu- the m/z 143 and 157 ions (which also indicates whether the tion mass spectrometry (HRMS) were performed on a Finnigan ion contains an odd or even number of electrons [10]), based MAT 95 HRMS with Finnigan MAT ICIS II operating sys- on elemental composition, results in a number of 1.5. This tem by Dr. Elliot Rachlin at the Mass Spectrometry Facility indicates that there is one ring or double bond in the ion struc- housed within the Department of Chemistry at the Univer- ture and that the ion under consideration is an even-electron sity of Utah. Samples were heated, starting from 0 to 20 ◦C ion. for 0.5 min, followed by additional heating from 40 ◦C/min to 280 ◦C, where the temperature was held for 10 min. Data were collected using electric sector scanning at a rate of 10 s/scan 2 For example, the default optimization settings for calculating the ener- (with a 0.4 s interscan time) over a m/z range of 129–171. The getic properties of water with Gaussian03 are accurate to within 0.001 kcal/mol single-scan spectrum with the greatest total ion current was [9]. C.R. Borges et al. / International Journal of Mass Spectrometry 247 (2005) 48–54 51

(6), a synthetic steroid with a structure identical to DHT with the 3.2. Molecular origin exception of the presence of a methyl group attached to carbon 2, were obtained. (The value of methylated analogs in the deter- It was observed that dihydrotestosterone (5), an endogenous mination of EI-fragment origins for TMS-derivatized steroids is steroid, also produces an ion with m/z 143 even though the steroid well founded [4].) The results showed the presence of a peak does not contain 17-hydroxy and 17-methyl functional groups. representing an ion with m/z 157 with the same elemental com- DHT was also analyzed by HRMS (data not shown) and it was position as the ion produced by ethylestrenol (4) supporting the discovered that the ion with m/z 143 produced by DHT has the A-ring origin hypothesis. Additional support for this hypothesis same elemental composition as that produced by mestanolone was gathered from the EI-mass spectrum of trimethylsilylated (2) and 3 -hydroxystanozolol (3). This means that some steroidal 2 [16,16,17- H3]-DHT (Fig. 3): the molecular ion was observed structural feature other than simultaneous 17-hydroxylation and at m/z 437 (as opposed to m/z 434 for native DHT), but no shift 17-methylation is capable of producing a very similar or identi- was observed in the peak at m/z 143 suggesting that the ion with cal ion to the one with m/z 143 that arises from synthetic steroids. m/z 143 arises from somewhere in the molecule other than the D- We hypothesized that this ion likely arose from the A-ring of ring. Apparently, under TMS-derivatized EI conditions, an ion DHT. To test this hypothesis HR mass spectra of dromostanolone with the same elemental composition is produced in abundance from the A-ring of 3-keto-enol steroids, as is produced from 17- hydroxy, 17-methyl steroids. Additional data (not shown) also demonstrates the generation of an ion (in low abundance, <25% RI) with m/z 143 from 3-hydroxy-4-ene steroids. Finally, Vouros and Harvey [11] have reported that an ion with the same elemental composition, but different structure than the ion of interest here, arises from a variety of 11-trimethylsiloxy steroids. These facts demonstrate that observation of an abnormally intense peak in the m/z 143 extracted ion chromatogram of urine samples undergoing screening for banned anabolic agents may not indicate the presence of a 17-hydroxy, 17-methyl steroid.

3.3. MS/MS

2 Fig. 3. EI-mass spectra of DHT (5) (Panel 1) and [16,16,17- H3]-DHT (Panel 2). The signiﬁcance of several important peaks is described in Sections 3.2 and To obtain detailed information on the structures of ions with 3.4. m/z 143 and 157 described in Sections 3.1 and 3.2, fragmentation 52 C.R. Borges et al. / International Journal of Mass Spectrometry 247 (2005) 48–54

at m/z 152 and 166, respectively, and had the expected relative abundance. (McCloskey et al. [15] have described the utility and interpretation of deuterium-labeled TMS derivatives in EI mass spectrometry.) Collision-induced dissociation studies on the deuterated ions with m/z 152 and m/z 166 were carried out, and analogs to products ions produced by the ions with m/z 143 and m/z 157 were observed with analogous abundances at m/z 82 (instead of m/z 73), m/z 100 (instead of m/z 91), and m/z 50 (instead of m/z 45) (data not shown). These speciﬁc m/z shifts support both the precursor- and the product-ion m/z assignments and structures. Fig. 4. Breakdown curve generated by CID and analysis by a second stage The breakdown curves obtained here unambiguously demon- of mass spectrometry for the major product ions of the m/z 143 precursor ion derived from TMS-derivatized epimetendiol (7). strate the production of trimethylsilyl and related ions upon CID of the ion with m/z 143. Many TMS-derivatized molecule frag- data at increasing collision energies were collected to construct ments produce the set of product ions observed in the MS/MS breakdown curves [12] of the major product ions of D- and A- spectra of the ion with m/z 143, but this cannot be assumed to ring derived ions with m/z 143 and of D- and A-ring derived occur without the proper evidence. Thus, by demonstrating the ions with m/z 157. Breakdown curves were obtained for epime- facile (very low CID energy) production of a trimethylsilyl ion, tendiol (7) (gives a D-ring, and possibly some A-ring derived the data generated lend support to the structure assigned to the ions with m/z 143), 3-hydroxystanozolol (3) (gives a D-ring ion with m/z 143 (9, Scheme 1). derived ion with m/z 143), mestanolone (2) (gives D- and A-ring derived ions with m/z 143), DHT (5) (gives an A-ring derived ion 3.4. Deuterium isotope labeling studies with m/z 143), 13␤,17␣-diethyl-3␣-17␤-dihydroxy-5␣-gonane (8) (gives a D-ring derived ion with m/z 157), ethylestrenol (4) As can be seen from a comparison of the mass spectra of (gives a D-ring derived ion with m/z 157), and dromostanolone DHT-d3 and native DHT (Fig. 3), a peak at m/z 131 constitutes (6) (gives an A-ring derived ion with m/z 157). For a given a2m/z unit increase over a peak at m/z 129 in the mass spec- MS/MS-derived product ion, a breakdown curve consists of a trum of native DHT. Based on previous studies on the EI-induced plot of the relative abundance of the product ion versus collision fragmentation of TMS-derivatized steroids [4], it is known that energy. When breakdown curves for several product ions are dis- the ion with m/z 129 in D-ring saturated, 17-hydroxysteroids includes carbon atoms C-17 and C-16. Thus, an increase of 2 m/z played together, information on relative ease of formation and 2 the identities of second, third, and subsequent generation prod- units in the m/z 129 ion of the [16,16,17- H3]-analog of DHT uct ions can be obtained [13]. Breakdown curves for all steroids rather than an increase of 3 m/z units is signiﬁcant because it analyzed were found to be nearly identical to that of epime- indicates that one of the hydrogen atoms attached to C-16 or C- tendiol (7)(Fig. 4) suggesting similar fragmentation pathways 17 is abstracted during formation of the ion. Based on analogy and structures for the D-ring and A-ring derived m/z 143 and to the well-characterized ethylene ketal EI-fragmentation path- m/z 157 ions. The breakdown curves indicate that the most eas- way [16,17], Diekman and Djerassi [4] suggest an explanation ily formed product ion fragment of the ion with m/z 143 and the for this 2 m/z unit shift in the form of a fragmentation pathway ion with m/z 157 is a trimethylsilyl ion (m/z 73) which gives rise (mechanism) by which hydrogen from C-16 is abstracted and to a methylsilyl ion (m/z 45) which dehydrogenates to produce formation of an ion with m/z 129 occurs. All experimental data an ion with m/z 43. (The methylsilyl ion has been documented to shown here, in agreement with these studies found in the litera- arise from the trimethylsilyl ion by others [14].) Alternate frag- ture, suggest that formation of the TMS-derivatized 17-hydroxy, mentation pathways give rise to a trimethylsilyloxonium ion (m/z 17-methyl steroid-derived ion with m/z 143 would form a highly 91) and a methylsilyloxonium ion (m/z 61). In all cases, the ion analogous product (9) (by difference of an electronically unin- with m/z 143 is one that easily produces, with less than 12 eV volved methyl group at C-17) via a highly analogous mechanism input energy, a trimethylsilyl ion upon CID. (Scheme 1). This mechanistic analogy is supported by Vouros and Harvey [18]. The ion structure (9) is identical to that initially suggested, but not experimentally proven, by Durbeck and Buker [5] after simple observation of an ion with m/z 143 in the spectra of TMS-derivatized 17-hydroxy, 17-methyl steroids. 2 Unfortunately, no [16,16- H2]-labeled 17-hydroxy, 17- methyl steroid is commercially available and we lack the facility and funds to either synthesize one ourselves or to commission a custom synthesis. Also, no deuterium-labeling studies were carried out to fully assess the structure and mechanism of formation Upon derivatization of steroids 2–8 with BSA-d18 (which of the A-ring derived ion with m/z 143 because such analogs were adds nine deuterium atoms per TMS group), the expected heavy not commercially available and because it is beyond the primary analogs of the ions with m/z 143 and m/z 157 were detected scope of the work presented here. C.R. Borges et al. / International Journal of Mass Spectrometry 247 (2005) 48–54 53

Scheme 1. Structure (9) and proposed mechanism of formation for the ion with m/z 143 derived from the D-ring of 17-hydroxy, 17-methyl steroids. The mechanism for facile formation of a trimethylsilyl ion (m/z 73) upon low energy CID of the m/z 143 ion is shown.

3.5. Stereochemistry Mulliken population analyses for the established ion structure (9) indicates that the positive charge is mostly located on the Si No major differences in abundance were observed for the atom (with an atomic partial charge greater than +0.8), whereas m/z 143 ion produced by epimetendiol (7) as compared to the partial charge of the oxygen atom is greater (more negative) that produced by 3-hydroxystanozolol (3) nor were differences than −0.4. These results point out that using chemical intu- observed for the ion with m/z 143 produced by oxandrolone as ition and formal charge assignments to put the positive charge compared to epioxandrolone, which are diastereomers (data not on a certain atom is not always accurate. The optimized struc- shown). Thus, stereochemistry appears not to play a signiﬁcant ture, geometrical parameters, and calculated infrared absorbance role in formation of the m/z 143 ion. spectrum of ion structure 9 were generated and are available upon request. 3.6. Quantum mechanical calculations 4. Conclusions Quantum mechanical calculations using density functional theory were used to establish the relative stabilities of proposed The results of this study verify that the ion with m/z 143 D-ring derived ion candidates that arose during the course of frequently observed in TMS-derivatized, EI-mass spectra of this investigation. As suggested by Gustaffson et al. [19] the 17-hydroxy, 17-methyl steroids arise from the D-ring of these production of highly stable end-point ions is a key driving force steroids. But, an ion with the same elemental composition for fragmentation pathways. This is true for unimolecular frag- (C7H15OSi) was also veriﬁed to arise from the A-ring of mentations, where the products with lower free energy would TMS-derivatized 3-keto-enol steroids such as DHT. Based on have the lower activation energy, in accord with the Hammond studies employing HRMS, GC/MS/MS, analog steroids, and postulate [20]. In terms of ground state free energy at 0 K (i.e., deuterium-labeling, the D-ring derived ion with m/z 143 was zero-point energy corrected potential energy), ion structure 9 determined to be a resonance stabilized, allylic ion, having a was found to be the most thermodynamically stable species (by 3-(O-trimethylsilyl)but-1-ene structure (9). The mechanism of 28.5 kcal/mol) amongst other candidates formed by initial het- formation for this ion is proposed (Scheme 1). erolytic fragmentation (not shown) that, ultimately, were found not to be in agreement with mass spectral evidence. This fact sug- Acknowledgements gests that these types of calculations may be valuable in guiding the acquisition of experimental (mass spectral) evidence during The authors express their thanks to Professor Michael D. the course of an investigation. Morse of the Department of Chemistry at the University of Utah 54 C.R. Borges et al. / International Journal of Mass Spectrometry 247 (2005) 48–54 for his helpful suggestions regarding some physical chemistry Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Strat- aspects of the work presented here. This work was primarily sup- mann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, ported by the NIH. Theoretical calculations were carried out with P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. the support of the NSF to TNT. LL acknowledges the Vietnamese Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Education Foundation for a graduate fellowship. Finnigan MAT Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. 95 work: acknowledgement is made to Dr. Elliot Rachlin of the Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, University of Utah’s Mass Spectrometry Facility, the National M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Science Foundation grant CHE-9002690, and to the University Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Revi- sion A.1 ed., Gaussian, Inc., Pittsburgh, PA, 2003. of Utah Institutional Funds Committee. [10] F.W. McLafferty, F. Turecek, Interpretation of Mass Spectra, University Science Books, Mill Valley, CA, 1993. References [11] P. Vouros, D.J. Harvey, J. Chem. Soc. (Perkin 1) 7 (1973) 727. [12] F.W. McLafferty, in: F.W. McLafferty (Ed.), Tandem Mass Spectrometry, [1] M. Saugy, C. Cardis, N. Robinson, C. Schweizer, Baillieres Best Pract. John Wiley and Sons, New York, 1983, p. 303. Res. Clin. Endocrinol. Metab. 14 (2000) 111. [13] R.D. Hiserodt, B.M. Pope, M. Cossette, M.L. Dewis, J. Am. Soc. Mass [2] B.G. Wolthers, G.P. Kraan, J. Chromatogr. A 843 (1999) 247. Spectrom. 15 (2004) 1462. [3] L.D. Bowers, Clin. Chem. 43 (1997) 1299. [14] A.E. Pierce, Silylation of Organic Compounds, Pierce Chemical Com- [4] J. Diekman, C. Djerassi, J. Org. Chem. 32 (1967) 1005. pany, Rockford, IL, 1968, p. 33. [5] H.W. Durbeck, I. Buker, Biomed. Mass Spectrom. 7 (1980) 437. [15] J.A. McCloskey, R.N. Stillwell, A.M. Lawson, Anal. Chem. 40 (1968) [6] C. Schoene, A.N. Nedderman, E. Houghton, Analyst 119 (1994) 233. 2537. [16] G. von Mutzenbecher, Z. Pelah, D.H. Williams, H. Budzikiewicz, [7] D.H. van de Kerkhof, R.D. van Ooijen, D. de Boer, R.H. Fokkens, N.M. Steroids 2 (1963) 475. Nibbering, J.W. Zwikker, J.H. Thijssen, R.A. Maes, J. Chromatogr. A [17] H. Budzikiewicz, C. Djerassi, D.H. Williams, Structure Elucidation of 954 (2002) 199. Natural Products by Mass Spectrometry, Holden-Day, San Francisco, [8] W. Van Thuyne, F.T. Delbeke, Biomed. Chromatogr. 18 (2004) 155. CA, 1964, p. 25. [9] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, [18] P. Vouros, D.J. Harvey, Anal. Chem. 45 (1973) 7. J.R. Cheeseman, J.A. Montgomery, T.V. Jr., K.N. Kudin, J.C. Burant, [19] J.A. Gustafsson, R. Ryhage, J. Sjovall, R.M. Moriarty, J. Am. Chem. J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Soc. 91 (1969) 1234. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M. Hada, [20] G.A. Hammond, J. Am. Chem. Soc. 77 (1955) 334. M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Naka- [21] L. Mateus-Avois, P. Mangin, M. Saugy, J. Chromatogr. B Analyt. Tech- jima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. nol. Biomed. Life Sci. 816 (2005) 193.